How not to measure temperature – part 29

concully1.jpg

The picture above is of Conconully, Washington and comes to me courtesy of Josiah Mault, of the Washington State Climate Office. Mault has been surveying all of the Washington stations for that office, and has been regularly making contributions to www.surfacestations.org The picture illustrates how human activity can spring up around a station. The MMTS electronic temperature sensor is shown next to a lean-to used for rafting gear storage. I presume the life preserver is placed next to the sensor as a reminder that we may need it in case of catastrophic sea level rise. The metal ore cart full of stones is a nice touch, and makes a perfect high mass IR radiative heatsource to keep the overnight lows a bit more “comfy”. There are also stones directly under the sensor whic you can see in this photo.

But perhaps it is not the curator’s fault, but rather that of the NWS/NOAA employee that made the placement, as we see in the next photo:
concully2.jpg
more pictures available here on surfacetstations.org

Once again, we have a climate station of record in the middle of a parking area, near buildings, and directly in the middle of regular human activity. One of the downsides to the NWS COOP modernization program started in the 1980’s and continuing today is the MMTS unit itself. It requires a cable, and that cable has be be buried to be brought into the domicile containing the electronic readout.

As anyone knows, especially rabbits, digging short holes is far easier than digging long ones. So its far easier and less time consuming to dig a short trench and place the sensor nearer the building. This proximity bias seems to have been repeated regularly when the MMTS system has replaced the traditional Stevenson Screen and Mercury Max-min thermometers.

There’s a reason that NOAA specifies that temperature sensors should be a minimum of 100 feet away from buildings, concrete, and asphalt which may introduce biases into the reading. What we don’t know is why there has been such an apparent regular failure to adhere to such specifications.

Advertisements

34 thoughts on “How not to measure temperature – part 29

  1. Perhaps the problem is not merely a lack of subterranean holesmanship- Does anyone know the length of cable that is supplied with the unit for its installation? Perhaps the supplier gives less than the required 100 plus feet!

  2. OK, I’ll try asking the question here:

    What exactly are the USHCN stations measuring?

    The temperature in 1200 places?

    The near average of temperatures in clear grassy, rural places across the US?

    The temperatures of a 4 to 5 ft shell of air across the lower 48?

    The trend of the above temps?

    Shouldn’t the system try to measure the air temp of the US weighted for altitude and land use as currently constituted (climate = Biosphere, Atmosphere, Lithosphere, Bathysphere)?

    If this is the case shouldn’t there be stations deep in the forests?

    In urban areas?

    In the middle of cropland?

    If the eventual expectation is that we will find a way to get some control over temperatures, what temperature should we be trying to control? Cropland, etc?

    Local or nationwide?

  3. I’d love to see the algorithm Dr. Hansen uses to “adjust” for a heat sink of large rocks in a metal container next to the MMTS unit.

  4. “If this is the case shouldn’t there be stations deep in the forests?

    “In urban areas?

    “In the middle of cropland?”

    If you didn’t want to have to adjust the data, theoretically you’d have to have the same proportion of urban sites as urban land area, the same % of cropland sites as there is cropland, etc.

    The problem here is that none of those things are stable over time, and the stations are not supposed to be moved.

    So we are more or less stuck with some sort of adjusted data.

    But, to paraphrase Mr. Gore, “all the adjustments that should be DOWN are UP and all the adjustments that should be UP are DOWN”.

    The surfacestations project is, in effect, an, “Adjust THIS, baby!” to the not-so-scientific community.

  5. O2, the answer to your question depends on what you are trying to measure. If you are trying to get an accurate measurement of a local area temperature for instantaneous use, then you want the station to be in an area that is most representative of the area where you need to know the temperature. On the other hand, if you are trying to see what the temperature change at a given location is over a period of many years, then the station needs to be placed in a location that is closest to the conditions under which the previous data were collected. That means if an area becomes urbanized, you are comparing apples to oranges. If you switch from measuring T in an open field to measuring it next to a building with a metal ore cart full of stones nearby, then you are comparing apples to oranges.

    The real devil is in the details of how you assign an adjustment to account for the presence of the building or the rocks or the idling vehicles in the driveway. The affects of each bias are likely to change over time, or be transient in nature, and as such will not make a clear pattern by which the magnitude of the affect will be readily apparent.

    The easy answer to me appears to be to figure out which stations suffer from the least potential bias and then use them to work backwards and figure out the impacts at the stations nearby with a more obvious bias. Hansen and others believe you can do it with a computer program, but they won’t tell us how the program works, won’t give access to the code, and won’t give any examples demonstrating how their adjustments account for the obvious biasing factors being found by folks like Anthony.

  6. You should avoid some of the grandstanding. You don’t yet know that this project will show anything and jumping up and down over individual stations becomes drama.

    Also, you don’t know that the stones will have a noticeable effect on observed night temps. You’re speculating. Also, the comment, “there’s a reason for the 100 foot distance” implies that you’ve seen proof of how effects of pavement and the like drop off with distance. But you have no idea. It could be that 10 feet is plenty…or that 1000 is needed. You just don’t know. So stop the drama. Be a fair scientist.

  7. Also, you don’t know that the stones will have a noticeable effect on observed night temps.

    TCO, you’ve never had a solar green house landscaped with boulders, stone or brick floor or water-filled 55 gal. drums, have you? Why do you supposed people put those kinds of things in greenhouses? I’ll give you a hint — it has something to do with the “night temps”.

  8. To TCO:

    Do you experience more cloudy days than those around you? Do small dark clouds follow you around?

    From what I’ve seen in the surfacestation photos, it’s mostly sunny everywhere in the US and it never seems to rain. Coincidence?

  9. Here I go again defending TCO. ( dude you owe me a great breakfast with bloody marys)

    TCO is a stickler for details. Yes, the spec calls for 100 ft clearence.

    The question is was any analysis done to support this? For example, someone may have studied this and found an effect out to 50 ft, and the spec ( be safe) is doubled. So, now
    comes the question, why 100 ft?

    On the other hand….

    When I see how sparsely sampled the rest of the world is, I’m inclined to TOSS every non compliant site in the US.

    TCO ( I trust I have this right friend) Thinks that Tossing this data is not justisfied, as of yet.

    So, I’m gunna go see if I can find any asphalt studies.

  10. Stuff: on asphalt

    http://www.cert.bham.ac.uk/research/urgent/Statisticalmodel.pdf

    Puts the question the other way, how does the surrundings cool the asphalt.. I suspect something could be teased out f this indirectly.

    interesting references and
    thermal mapping studies

    Thornes, J.E. (1991). Thermal mapping and road-weather information systems for highway
    engineers, in Perry, A.H. & Symons, L.J. (eds) Highway Meteorology, E & FN Spon,
    London., 39-67pp
    Thornes, J.E. & Shao, J. (1991a). A comparison of UK ice prediction models. Meteorol.l
    Mag., 120: 51-57.
    Thornes, J.E. & Shao, J. (1991b). Spectral analysis and sensitivity tests for a numerical road
    surface temperature prediction model. Meteorol. Mag., 120: 117-123.

    Also, this

    http://trb.metapress.com/content/d1v0430412hw4762/

    http://www.treelink.org/joa/2004/jul/celestian7-04.pdf

    From the latter.

    But impervious surfaces around parking lot
    median landscapes such as asphalt and concrete absorb,
    store, and reradiate more heat energy per unit area than do
    vegetated surface covers such as turfgrass, which is able to
    dissipate heat energy through evapotranspiration of water
    (Holman 1986). As a result, air, surface, and soil temperatures
    near asphalt and concrete might be higher than
    temperatures in the near vicinity of vegetated surfaces
    (Asaeda et al. 1996; Montague et al. 2000).
    In the past, some researchers have studied thermal
    microenvironments that affect the growth and performance
    of landscape trees. Kjelgren and Montague (1998) recorded
    asphalt surface temperatures that were as much as 20°C to
    25°C (36°F to 45°F) higher than surrounding surfaces
    landscaped with vegetation. Graves and Dana (1987)
    reported that mean summer (July) soil temperatures at
    depths of 5 to 50 cm (2 to 20 in.) under street trees in
    downtown Lafayette, Indiana, U.S., were up to 7°C (13°F)
    higher than on urban and rural sites not surrounded by
    paved surfaces. In central California, U.S., Scott et al. (1999)
    found that parking lot surfaces shaded by trees were as
    much as 20°C (36°F) lower than adjacent sunlit surfaces
    during July and August.

  11. Clayton: Thanks.

    It didn’t specify raw or adjusted, but I will assume it’s adjusted.

    And what do w see?

    The usual trend up since 1980, which fits in with my as-yet-unfounded prejudices (TBD). And a bit of a dip around 1940, though, unlike the US and world trend.

    Makes one really want to see what trends the raw data for the type 1 surface stations show once they’ve all been observed.

    OTOH, all the speculation is meaningless: The results will be what they will be. Just a matter of patience.

  12. Sounds to me, gentlemen, that a new experiment is called for. I don’t see why one would be content with inference. (Besides, that’s how Dr. Hansen got in trouble in the first place. He inferred too much.)

    Someone who knows how (I.e., not me) should construct a controlled experiment to measure the effet of specific violations (plus various combos of violations) to be compared with a nearby well-sited example.

    Trouble is, those stations cost what? A thou a throw? Even a few is running into serious bucks. Not to mention time, gas, bother, and misc. expenses.

    Of course, the Rev has three stations already (purchased, I believe, at his own expense): he’s using them for his latex Vs Whitewash experiment. Perhaps when that is done, they could be used to assess–by actual measurement–the specific effect of the various microsite violations.

  13. From what I’ve seen in the surfacestation photos, it’s mostly sunny everywhere in the US and it never seems to rain. Coincidence?

    Well, I suspect that an aversion to subjecting digital cameras (and surveyers) to the rain might introduce a systemic bias to that observation.

  14. Anthony, when you say “there’s a reason for the 100 foot spec”, just citing and proving that it is in fact “a spec” is not sufficient. You need to show “there’s a reason”. IOW, why 100 vice 400 or vice 25?

  15. Not sufficient for whom, TCO? Anthony didn’t write the specs, someone at NOAA did. Perhaps whoever wrote the specs just thought 100 feet sounded like a reasonable number. Or perhaps they spent millions of your (and my) tax dollars in exhaustive experimentation to determine the optimum distance.

  16. TCO,

    I already said that for you. And I said it better.

    Go read some of the papers I have already Posted and help nail this thing down. There is tons more stuff but Not in the usual places

  17. TCO,

    I cant find this online..

    Do you access to a library

    Leroy, M. Meteorological Measurement Representativity, Nearby Obstacles Influence, 10th Symposium on Meteorological Observations and Instrumentation, American Meteorological Society, January 11-16, 1998.

  18. Retired Spook,

    I’m sure it’s the former rather than the latter. In my experience, anytime you see a nice round number like 100 in a spec, its a WAG. Now if it said something like 65 feet for concrete, 90 feet for asphalt I’d be more inclined to think someone did some sort of an analysis or test to come up with the numbers.

    All that being said, a spec is a spec, and it’s being routinely violated. If NWS has good reason to believe the 100 foot spacing isn’t needed, then change (or delete) the spec, don’t just pretend it doesn’t exist. That’s just good engineering practice.

  19. I was standing underneath a bank of four ac units mounted on a medical office building today. The ambient temperature outdoors was 93 ~ 2 degrees.
    I could feel extra heat from 30 feet away.
    The units were mounted 7 feet off the ground.

    You we don’t have to use an elaborate NOAA approved temperature set up, to make the point that there is a difference between a thermometer in a rock bed and one in a grass field.

    Just 2 or 3 of the same type of store bought 5 dollar mercury thermometers should do it.

  20. TCO, I can’t speak for NOAA/NWS as to why they chose that number any more than any one of us could speak for why we obey speed limits on roads. It’s a spec, the spec is accepted, its followed.

    If you don’t follow the spec, trouble follows. Operating outside of the spec has risks, be it a ticket or faulty data.

  21. Retired Spook, I’m sure it’s the former rather than the latter. In my experience, anytime you see a nice round number like 100 in a spec, its a WAG.

    Jeff C., I was sorta jerking TCO’s chain. From Anthony’s latest post, it appears that it may actually have been the latter, and done with French tax dollars, no less — a twofer. It’s good to know that some thought went into the specs, even if so many stations are ignoring them.

  22. Steve: thanks for your cogent statements and the supply of papers.

    Anthony: Thanks for your statement. I understand your, “it’s the law, I didn’t write the law, but we should follow it” standpoint. That’s all well and good. It’s just when you make rhetorical flourishes, along the lines of “it’s the law and there’s a damn good reason for that law”. Well, then you need to show the damn good reason for it. Not just rever to “the law is the law”. (Not trying to be pedantic, but I’m still not clear if you track on this basic concept that both Mosh and myself have described.)

  23. “Perhaps the supplier gives less than the required 100 plus feet!”

    That may be, but if you look at the photos for Newberry MI, you’ll see they overcame it. That certainly appears to be close to the distance spec.

  24. TCO,

    I have not found anything directly explaining what I would call the 4D thermal plume from a heat storage device. Simply, how does the heat flow over space and time?

    Being reasonable fellows we agree that asphalt, concrete and rock act like capacitors and store heat. And they store it differently than grass, dirt, and dead squirrels.

    When these capacitor’s release we have heat flow Houston. The question is..

    what the heck does that flow feild look like?
    And what does it look like
    over time?

    I Suspect the answer is NOT 49. I suppose that answer is as hard as the chinese remainder thereom.

    Now, the Mandelbrotian within me says this.
    Understanding the Microsite problem IS NO DIFFERENT than understanding the Global Climate problem.

    It’s turtles all the way down..

    To other folks . You want someone like TCO pushing against you. He has never questioned my politics or motivation and has only put my belief under the microscope.
    So, have a good yell at him if you must ( I did) but don’t disregard his prodding.

  25. It’s not just a capacitor effect. It’s direct warming from radiation capture during the day (for instance a fresh-laid asphalt blacktop being near worst), versus reflection.

    I have no idea how severe the problem is. Neither does Anthony. I’m not impressed by his gut though. For instance with the air conditioners, he was way to quick to assume very strong effects from distant, second story window units.

  26. And I’m not impressed with TCO’s circular logic.

    You used the words “very large effect” in describing my assessment of 2nd story a/c.

    Now prove I said those exact words. I don’t believe I did.v

    a/c units can have an effect based on wind conditions, local structure and foliage ducting of wind, eddies, and the a/c output. For example, Happy Camp CA 1st and 2nd story units can have such effects from SW winds.

    But thats not the point, and you seem to keep missing it.

    The point is this: stations used to monitor climate shouldn’t be in proximity with such heat sources in the first place, or be near buildings, or parking lots, etc because to disentangle all the potential biases, inevitable leads to attempts to quantify them. And it may not be possible.

    Gavin says we need 60 good sites, we should identify them and see what trend they have compared to CONUS and ROW.

    BTW TCO, how about a name? I don’t like debating phantoms….like Eli, BCL, Tamino, and Dano whom hide behind monikers. If yoiu believe enough in what you have to say, put you name to it…thats my world view.

    Imagine how far Hansen would have gotten with a handle instead of a name.

Comments are closed.